(present: TI, TH, RK, ML, MF, AA)

Tamara investigated some the remaining mysteries of the E(hec) vs E(emec) plots. There are four distinct regions in the plot:
1) A diagonal band, counter-correlated between E(hec) and E(emec). These are pions that shower early.
2) A big cluster of events with E(emec) around zero and E(hec) large. (Those are pions too - about 30% of all events.)
3) A thin band of events with E(emec) around zero and E(hec) significant, although not as large as the previous cluster. (These needed to be explained!)
4) Events with E(hec) and E(emec) around zero. (Those are almost certainly muons.)
Tamara made sure that the timing was ok for the events she chose, and then she looked at the energy deposited in HEC. If there is any significant energy measured there, it will most likely be due to a shower that started late. I turns out at a large fraction of the mysterious events of group 3) fit that hypothesis. Simple calculations show that the expected number of pions that would not shower in the EMEC is about 25%, and this is consistent with the number of events in grouo 3). Tamara and Michel also calculated that there are about 8.5 interaction lengths total to the end of the calorimeter, which corresponds to a 95% containment. The `missing' 5% are consistent with the Monte Carlo estimates.
Tamara has now started to work on timing corrections. She takes only the central 22 ns out of the timing distribution. This improves the response minutely. Her next aim is to fit the distribution and to perform timing corrections at the cluster level.
Michel pointed out that the details of timing corrections are not trivial - it is not clear how to do the timing correction such that impact point to impact point relative normalization for the runs is conserved.

Tayfun varied the power of his energy weights n (E^n) and looked what effect it has on finding the position in the EMEC. In the x-direction, a power of 1 looks fine for all 3 layers. In the y direction, a power of 1 is ok for the first layer, but the second layer favours two spots: either 0.5 or something around 1.5-2. The third layer is inconclusive - anywhere between n=0.5 and n=2.
Richard pointed out that he remembers from Opal analyses that n=0.5 was favoured when the cluster size is less that the cell size and larger n was favoured if the cluster size is larger than the cell size.
Margret then surmised that the y-effect in the second layer is likely due to the slanting cell structure in our testbeam-setup. Depending on the beam energy and resulting shower depth more or less cells get traversed. In the third layer the cells are relatively large, and there it really depends on whether the shower hits one cell in the middle or whether it hits in-between two cells, sharing the energy deposition between them.